Anthropic CEO Dario Amodei has resumed negotiations with the Pentagon following a public dispute regarding AI safety protocols. The Department of Defense previously demanded the removal of safeguards preventing the Claude AI model from performing mass domestic surveillance or operating fully autonomous weapons.
Anthropic refused the demand on ethical grounds, prompting the government to threaten the cancellation of a $200 million contract. Federal officials subsequently labeled the startup a supply chain risk.
Amodei confirmed the de-escalation efforts during a Morgan Stanley conference. Rival firm OpenAI secured a separate Pentagon contract during the initial breakdown of these talks.
The renewed discussions indicate a potential compromise for deploying advanced AI in national security contexts while preserving ethical guardrails.